An apparatus for recording and reproducing digital audio signals (referred to as PCM signals) by a pair of rotary heads is disclosed in, for example, U.S. Pat. No. 4,523,237 by the same applicant as the present invention. In that apparatus an error correction code is used such that the PCM signals of the amount to be recorded and reproduced by one rotary head are arranged in a matrix form, an error correction coding process C.sub.1 is performed for each PCM signal arranged in the vertical direction of the matrix arrangement, and an error correction coding process C.sub.2 is executed for each PCM signal arranged in the lateral direction of the matrix arrangement. The PCM signal which was subjected to the error correction coding process and the vertically arranged check symbol of the error correction code C.sub.1 is recorded and reproduced as a unit. Similarly, the PCM signal which was subjected to the error correction coding process and the vertically arranged check symbol of the error correction code C.sub.2 is also recorded and reproduced as a unit. To error correct the reproduced signal, decoding of the error correction code C.sub. 1 (C.sub.1 decoding) is performed and the decoding of the error correction code C.sub.2 (C.sub.2 decoding) is then executed.
When performing the C.sub.1 and C.sub.2 decoding, one matrix and the corresponding reproduced data are written into the buffer memory. Next, the data read out of the buffer memory for each series of the code C.sub.1 is subjected to the error correcting process and a corrected data and the pointer indicative of the presence or absence of errors are written into the buffer memory (C.sub.1 decoding). The data which is stored in the buffer memory and was subjected to the C.sub.1 decoding and the pointer are read out for each series of the code C.sub.2, the data which was read out is subjected to the error correcting process, and the corrected data and the pointer are written into the buffer memory (C.sub.2 decoding). The PCM data which was subjected to C.sub.1 and C.sub.2 decoding and the pointers are finally read out of the buffer memory in accordance with the inherent order of the data and the error data is interpolated by an interpolating circuit in a known manner.
In the foregoing decoding processes, after the reproduced data was once stored into the buffer memory, the C.sub.1 decoding and C.sub.2 decoding are executed. However, it takes a predetermined period of time to perform the C.sub.1 and C.sub.2 decoding. If the decoding is performed after the reproduced data was once stored into the buffer memory, the decoding cannot be sufficiently executed.